Jerry Fochtman <[log in to unmask]> wrote (in part)


<snip>

>This indeed is the dilemma, for one risks breaking forward compatibility
>without any type of indication to the user/process.  Given that forward
>compatibility is one of the HP3K's strong points, I would agree with the
>approach to indicate a failure in the field and supply a new LISTF, mode
>for the large files.  In this manner, should a current production process
>encounter one it should 'break', and then it can be updated for the new
>format.  Otherwise the process may not 'break', proceeding with erroneous
>results.
>

I agree with Jerry. Forward compatibility is one of many features that
distinguishes MPE from other operating systems. Another is each and every
system behaves the same way to standard MPE commands. Some of the
suggestions that have been proposed makes MPE more like Unix every day. Some
of the things I don't care for are:

o Global switches/variables that control how the output of listf looks
o "To h*ll with the users. If they want to go to the next version let them
fix their scripts/programs/jobs"

MPE has always taken the safe approach. When in doubt and data could be
corrupted the system will abort. Why then would you not want a script to
fail and a job terminate immediately if there is a risk of bad data?

Doug Werth                                     Beechglen Development Inc.
[log in to unmask]                                       Cincinnati, Ohio

The opinions expressed do not necessarily represent the views or opinions
of Beechglen Development. They might, but not necessarily. They represent
solely the opinions of the author.